AI automation for media buying on Claude Code: UGC creatives in batches, reports and campaign alerts with no manual work. Plus UA consulting. In performance since 2014.
A competitor’s winning ad → 5 variations of your own, 9:16. Your time: 15 minutes for three approvals.
See the case →A summary with takeaways in Slack every Monday. GitHub Actions, open source.
Code and details →CPI, ROAS and spend: a Slack alert with the reason, not just a number.
Code and details →For media buyers and UA teams. Every pipeline adapts to your stack: AppsFlyer, Adjust, Meta, TikTok, spreadsheets and Slack.
GitHub Actions runs them on cron and posts to Slack. A failure doesn’t break the report: it still goes out on time, just without the conclusions block.
We take an ad that already pays back for another product and carry over its mechanics. The words and shots are new; your app’s interface is real.
mainmain versionvoice_bsecond voiceproduct_hookhook: product screentext_hooktext hook on topshort~11 s: hook → proof → CTAEach variation differs from the main one in a single parameter, so the test shows exactly what worked: the voice, the first frame or the length.
Weekfork is a fictional meal-plan service. Pick an ad and the page in the phone rebuilds around its promise and offer.
One page, 4 angles × 3 offers = 12 versions. The version comes from the ad name — no manual setup per creative.
This is a simplified demo of my Soft Launch Advisor. The full version with benchmark reasoning and budget debates — on a call.
In performance since 2014, budgets up to $500K/mo in Tier-1, experience with Meta, TikTok, Snapchat, Google Ads, ASA, in-app networks.
Countries, sources, KPIs and test plan — so you don't burn budget on guesses.
Meta, TikTok, Google Ads, Snapchat, in-app networks — channel selection for your product and unit economics.
A workaround for SKAN limitations that's already boosting metrics for tier-1 apps.
Profitable markets, risks, path from $50K to $500K/mo without breaking unit economics.
What to test, what's burning out, which format is underused — without a month of trial and error.
Which features and metrics to move so marketing budget works instead of leaking.
Started in 2014 as a media buyer. Went from running first campaigns to managing teams and budgets in Tier-1 markets. Over the last 2 years — 5 products scaled from zero to $1M ARR. Current focus: AI automations for UA teams — building Claude Code pipelines so routine work disappears and the team can focus on strategy.
Picking countries, sources and a test plan so the first $30–50K don't go to guesses. I know which sources give signal fast and which ones quietly burn budget.
When Meta and Google hit a ceiling — Liftoff, Mintegral, Applovin, Moloco, OEM (Vivo, Oppo, Huawei) open up. Experience with this stack on $500K+/mo budgets.
A workaround for SKAN limits that boosts iOS campaigns. Web-sub, web2app, web2web — built on real projects, not just read in blogs.
Briefs for designers and video production, format testing, collaboration with product. I know how to run a creative pipeline, not "wait a week for one banner".
From a 1-hour call to full UA management. Starting prices are public — no subscriptions or hidden fees.
I take the creative that already works for you and build a series of variations to test. We test what works, not random ideas.
Review of your UA strategy, campaign audit, or automation advice. Output: a list of concrete actions, not generic words.
I'll build one pipeline from the section above for your stack. Or we'll design a custom one — for a routine that eats 5+ hours per week.
Launching and scaling UA — strategy, campaigns, creatives, reporting. Good when you need an external lead instead of hiring in-house.
20 minutes free call: tell me about your product and current tasks, I'll say what I can actually help with — and where I won't save you a single hour. No pitch, no sales.